namespace X86ISD {
// X86 Specific DAG Nodes
- enum NodeType {
+ enum NodeType : unsigned {
// Start the numbering where the builtin ops leave off.
FIRST_NUMBER = ISD::BUILTIN_OP_END,
- /// BSF - Bit scan forward.
- /// BSR - Bit scan reverse.
+ /// Bit scan forward.
BSF,
+ /// Bit scan reverse.
BSR,
- /// SHLD, SHRD - Double shift instructions. These correspond to
+ /// Double shift instructions. These correspond to
/// X86::SHLDxx and X86::SHRDxx instructions.
SHLD,
SHRD,
- /// FAND - Bitwise logical AND of floating point values. This corresponds
+ /// Bitwise logical AND of floating point values. This corresponds
/// to X86::ANDPS or X86::ANDPD.
FAND,
- /// FOR - Bitwise logical OR of floating point values. This corresponds
+ /// Bitwise logical OR of floating point values. This corresponds
/// to X86::ORPS or X86::ORPD.
FOR,
- /// FXOR - Bitwise logical XOR of floating point values. This corresponds
+ /// Bitwise logical XOR of floating point values. This corresponds
/// to X86::XORPS or X86::XORPD.
FXOR,
- /// FANDN - Bitwise logical ANDNOT of floating point values. This
+ /// Bitwise logical ANDNOT of floating point values. This
/// corresponds to X86::ANDNPS or X86::ANDNPD.
FANDN,
- /// FSRL - Bitwise logical right shift of floating point values. These
- /// corresponds to X86::PSRLDQ.
- FSRL,
-
- /// CALL - These operations represent an abstract X86 call
+ /// These operations represent an abstract X86 call
/// instruction, which includes a bunch of information. In particular the
/// operands of these node are:
///
///
CALL,
- /// RDTSC_DAG - This operation implements the lowering for
- /// readcyclecounter
+ /// This operation implements the lowering for readcyclecounter
RDTSC_DAG,
/// X86 Read Time-Stamp Counter and Processor ID.
/// 1 is the number of bytes of stack to pop.
RET_FLAG,
- /// REP_STOS - Repeat fill, corresponds to X86::REP_STOSx.
+ /// Repeat fill, corresponds to X86::REP_STOSx.
REP_STOS,
- /// REP_MOVS - Repeat move, corresponds to X86::REP_MOVSx.
+ /// Repeat move, corresponds to X86::REP_MOVSx.
REP_MOVS,
- /// GlobalBaseReg - On Darwin, this node represents the result of the popl
+ /// On Darwin, this node represents the result of the popl
/// at function entry, used for PIC code.
GlobalBaseReg,
- /// Wrapper - A wrapper node for TargetConstantPool,
+ /// A wrapper node for TargetConstantPool,
/// TargetExternalSymbol, and TargetGlobalAddress.
Wrapper,
- /// WrapperRIP - Special wrapper used under X86-64 PIC mode for RIP
+ /// Special wrapper used under X86-64 PIC mode for RIP
/// relative displacements.
WrapperRIP,
- /// MOVDQ2Q - Copies a 64-bit value from the low word of an XMM vector
+ /// Copies a 64-bit value from the low word of an XMM vector
/// to an MMX vector. If you think this is too close to the previous
/// mnemonic, so do I; blame Intel.
MOVDQ2Q,
- /// MMX_MOVD2W - Copies a 32-bit value from the low word of a MMX
+ /// Copies a 32-bit value from the low word of a MMX
/// vector to a GPR.
MMX_MOVD2W,
- /// PEXTRB - Extract an 8-bit value from a vector and zero extend it to
+ /// Copies a GPR into the low 32-bit word of a MMX vector
+ /// and zero out the high word.
+ MMX_MOVW2D,
+
+ /// Extract an 8-bit value from a vector and zero extend it to
/// i32, corresponds to X86::PEXTRB.
PEXTRB,
- /// PEXTRW - Extract a 16-bit value from a vector and zero extend it to
+ /// Extract a 16-bit value from a vector and zero extend it to
/// i32, corresponds to X86::PEXTRW.
PEXTRW,
- /// INSERTPS - Insert any element of a 4 x float vector into any element
+ /// Insert any element of a 4 x float vector into any element
/// of a destination 4 x floatvector.
INSERTPS,
- /// PINSRB - Insert the lower 8-bits of a 32-bit value to a vector,
+ /// Insert the lower 8-bits of a 32-bit value to a vector,
/// corresponds to X86::PINSRB.
PINSRB,
- /// PINSRW - Insert the lower 16-bits of a 32-bit value to a vector,
+ /// Insert the lower 16-bits of a 32-bit value to a vector,
/// corresponds to X86::PINSRW.
PINSRW, MMX_PINSRW,
- /// PSHUFB - Shuffle 16 8-bit values within a vector.
+ /// Shuffle 16 8-bit values within a vector.
PSHUFB,
- /// ANDNP - Bitwise Logical AND NOT of Packed FP values.
+ /// Compute Sum of Absolute Differences.
+ PSADBW,
+ /// Compute Double Block Packed Sum-Absolute-Differences
+ DBPSADBW,
+
+ /// Bitwise Logical AND NOT of Packed FP values.
ANDNP,
- /// PSIGN - Copy integer sign.
+ /// Copy integer sign.
PSIGN,
- /// BLENDV - Blend where the selector is a register.
- BLENDV,
-
- /// BLENDI - Blend where the selector is an immediate.
+ /// Blend where the selector is an immediate.
BLENDI,
- /// ADDSUB - Combined add and sub on an FP vector.
+ /// Blend where the condition has been shrunk.
+ /// This is used to emphasize that the condition mask is
+ /// no more valid for generic VSELECT optimizations.
+ SHRUNKBLEND,
+
+ /// Combined add and sub on an FP vector.
ADDSUB,
- // SUBUS - Integer sub with unsigned saturation.
+ // FP vector ops with rounding mode.
+ FADD_RND,
+ FSUB_RND,
+ FMUL_RND,
+ FDIV_RND,
+ FMAX_RND,
+ FMIN_RND,
+ FSQRT_RND,
+
+ // FP vector get exponent
+ FGETEXP_RND,
+ // Extract Normalized Mantissas
+ VGETMANT,
+ // FP Scale
+ SCALEF,
+ // Integer add/sub with unsigned saturation.
+ ADDUS,
SUBUS,
-
- /// HADD - Integer horizontal add.
+ // Integer add/sub with signed saturation.
+ ADDS,
+ SUBS,
+ // Unsigned Integer average
+ AVG,
+ /// Integer horizontal add.
HADD,
- /// HSUB - Integer horizontal sub.
+ /// Integer horizontal sub.
HSUB,
- /// FHADD - Floating point horizontal add.
+ /// Floating point horizontal add.
FHADD,
- /// FHSUB - Floating point horizontal sub.
+ /// Floating point horizontal sub.
FHSUB,
- /// UMAX, UMIN - Unsigned integer max and min.
- UMAX, UMIN,
+ // Integer absolute value
+ ABS,
- /// SMAX, SMIN - Signed integer max and min.
- SMAX, SMIN,
+ // Detect Conflicts Within a Vector
+ CONFLICT,
- /// FMAX, FMIN - Floating point max and min.
- ///
+ /// Floating point max and min.
FMAX, FMIN,
- /// FMAXC, FMINC - Commutative FMIN and FMAX.
+ /// Commutative FMIN and FMAX.
FMAXC, FMINC,
- /// FRSQRT, FRCP - Floating point reciprocal-sqrt and reciprocal
- /// approximation. Note that these typically require refinement
+ /// Floating point reciprocal-sqrt and reciprocal approximation.
+ /// Note that these typically require refinement
/// in order to obtain suitable precision.
FRSQRT, FRCP,
- // TLSADDR - Thread Local Storage.
+ // Thread Local Storage.
TLSADDR,
- // TLSBASEADDR - Thread Local Storage. A call to get the start address
+ // Thread Local Storage. A call to get the start address
// of the TLS block for the current module.
TLSBASEADDR,
- // TLSCALL - Thread Local Storage. When calling to an OS provided
+ // Thread Local Storage. When calling to an OS provided
// thunk at the address from an earlier relocation.
TLSCALL,
- // EH_RETURN - Exception Handling helpers.
+ // Exception Handling helpers.
EH_RETURN,
- // EH_SJLJ_SETJMP - SjLj exception handling setjmp.
+ // SjLj exception handling setjmp.
EH_SJLJ_SETJMP,
- // EH_SJLJ_LONGJMP - SjLj exception handling longjmp.
+ // SjLj exception handling longjmp.
EH_SJLJ_LONGJMP,
- /// TC_RETURN - Tail call return. See X86TargetLowering::LowerCall for
+ /// Tail call return. See X86TargetLowering::LowerCall for
/// the list of operands.
TC_RETURN,
- // VZEXT_MOVL - Vector move to low scalar and zero higher vector elements.
+ // Vector move to low scalar and zero higher vector elements.
VZEXT_MOVL,
- // VZEXT - Vector integer zero-extend.
+ // Vector integer zero-extend.
VZEXT,
- // VSEXT - Vector integer signed-extend.
+ // Vector integer signed-extend.
VSEXT,
- // VTRUNC - Vector integer truncate.
+ // Vector integer truncate.
VTRUNC,
+ // Vector integer truncate with unsigned/signed saturation.
+ VTRUNCUS, VTRUNCS,
- // VTRUNC - Vector integer truncate with mask.
- VTRUNCM,
-
- // VFPEXT - Vector FP extend.
+ // Vector FP extend.
VFPEXT,
- // VFPROUND - Vector FP round.
+ // Vector FP round.
VFPROUND,
- // VSHL, VSRL - 128-bit vector logical left / right shift
+ // Vector signed/unsigned integer to double.
+ CVTDQ2PD, CVTUDQ2PD,
+
+ // 128-bit vector logical left / right shift
VSHLDQ, VSRLDQ,
- // VSHL, VSRL, VSRA - Vector shift elements
+ // Vector shift elements
VSHL, VSRL, VSRA,
- // VSHLI, VSRLI, VSRAI - Vector shift elements by immediate
+ // Vector shift elements by immediate
VSHLI, VSRLI, VSRAI,
- // CMPP - Vector packed double/float comparison.
+ // Vector packed double/float comparison.
CMPP,
- // PCMP* - Vector integer comparisons.
+ // Vector integer comparisons.
PCMPEQ, PCMPGT,
- // PCMP*M - Vector integer comparisons, the result is in a mask vector.
+ // Vector integer comparisons, the result is in a mask vector.
PCMPEQM, PCMPGTM,
- /// CMPM, CMPMU - Vector comparison generating mask bits for fp and
+ /// Vector comparison generating mask bits for fp and
/// integer signed and unsigned data types.
CMPM,
CMPMU,
+ // Vector comparison with rounding mode for FP values
+ CMPM_RND,
- // ADD, SUB, SMUL, etc. - Arithmetic operations with FLAGS results.
+ // Arithmetic operations with FLAGS results.
ADD, SUB, ADC, SBB, SMUL,
INC, DEC, OR, XOR, AND,
- BEXTR, // BEXTR - Bit field extract
+ BEXTR, // Bit field extract
UMUL, // LOW, HI, FLAGS = umul LHS, RHS
- // MUL_IMM - X86 specific multiply by immediate.
+ // 8-bit SMUL/UMUL - AX, FLAGS = smul8/umul8 AL, RHS
+ SMUL8, UMUL8,
+
+ // 8-bit divrem that zero-extend the high result (AH).
+ UDIVREM8_ZEXT_HREG,
+ SDIVREM8_SEXT_HREG,
+
+ // X86-specific multiply by immediate.
MUL_IMM,
- // PTEST - Vector bitwise comparisons.
+ // Vector bitwise comparisons.
PTEST,
- // TESTP - Vector packed fp sign bitwise comparisons.
+ // Vector packed fp sign bitwise comparisons.
TESTP,
- // TESTM, TESTNM - Vector "test" in AVX-512, the result is in a mask vector.
+ // Vector "test" in AVX-512, the result is in a mask vector.
TESTM,
TESTNM,
// OR/AND test for masks
KORTEST,
+ KTEST,
// Several flavors of instructions with vector shuffle behaviors.
PACKSS,
PSHUFHW,
PSHUFLW,
SHUFP,
+ //Shuffle Packed Values at 128-bit granularity
+ SHUF128,
MOVDDUP,
MOVSHDUP,
MOVSLDUP,
MOVSS,
UNPCKL,
UNPCKH,
- VPERMILP,
+ VPERMILPV,
+ VPERMILPI,
VPERMV,
VPERMV3,
VPERMIV3,
VPERMI,
VPERM2X128,
+ //Fix Up Special Packed Float32/64 values
+ VFIXUPIMM,
+ //Range Restriction Calculation For Packed Pairs of Float32/64 values
+ VRANGE,
+ // Reduce - Perform Reduction Transformation on scalar\packed FP
+ VREDUCE,
+ // RndScale - Round FP Values To Include A Given Number Of Fraction Bits
+ VRNDSCALE,
+ // VFPCLASS - Tests Types Of a FP Values
+ VFPCLASS,
+ // Broadcast scalar to vector
VBROADCAST,
- // masked broadcast
- VBROADCASTM,
+ // Broadcast subvector to vector
+ SUBV_BROADCAST,
// Insert/Extract vector element
VINSERT,
VEXTRACT,
- // PMULUDQ - Vector multiply packed unsigned doubleword integers
+ /// SSE4A Extraction and Insertion.
+ EXTRQI, INSERTQI,
+
+ // Vector multiply packed unsigned doubleword integers
PMULUDQ,
- // PMULUDQ - Vector multiply packed signed doubleword integers
+ // Vector multiply packed signed doubleword integers
PMULDQ,
-
+ // Vector Multiply Packed UnsignedIntegers with Round and Scale
+ MULHRS,
+ // Multiply and Add Packed Integers
+ VPMADDUBSW, VPMADDWD,
// FMA nodes
FMADD,
FNMADD,
FNMSUB,
FMADDSUB,
FMSUBADD,
-
- // VASTART_SAVE_XMM_REGS - Save xmm argument registers to the stack,
- // according to %al. An operator is needed so that this can be expanded
- // with control flow.
+ // FMA with rounding mode
+ FMADD_RND,
+ FNMADD_RND,
+ FMSUB_RND,
+ FNMSUB_RND,
+ FMADDSUB_RND,
+ FMSUBADD_RND,
+
+ // Compress and expand
+ COMPRESS,
+ EXPAND,
+
+ //Convert Unsigned/Integer to Scalar Floating-Point Value
+ //with rounding mode
+ SINT_TO_FP_RND,
+ UINT_TO_FP_RND,
+
+ // Vector float/double to signed/unsigned integer.
+ FP_TO_SINT_RND, FP_TO_UINT_RND,
+ // Save xmm argument registers to the stack, according to %al. An operator
+ // is needed so that this can be expanded with control flow.
VASTART_SAVE_XMM_REGS,
- // WIN_ALLOCA - Windows's _chkstk call to do stack probing.
+ // Windows's _chkstk call to do stack probing.
WIN_ALLOCA,
- // SEG_ALLOCA - For allocating variable amounts of stack space when using
+ // For allocating variable amounts of stack space when using
// segmented stacks. Check if the current stacklet has enough space, and
// falls back to heap allocation if not.
SEG_ALLOCA,
- // WIN_FTOL - Windows's _ftol2 runtime routine to do fptoui.
- WIN_FTOL,
-
// Memory barrier
MEMBARRIER,
MFENCE,
SFENCE,
LFENCE,
- // FNSTSW16r - Store FP status word into i16 register.
+ // Store FP status word into i16 register.
FNSTSW16r,
- // SAHF - Store contents of %ah into %eflags.
+ // Store contents of %ah into %eflags.
SAHF,
- // RDRAND - Get a random integer and indicate whether it is valid in CF.
+ // Get a random integer and indicate whether it is valid in CF.
RDRAND,
- // RDSEED - Get a NIST SP800-90B & C compliant random integer and
+ // Get a NIST SP800-90B & C compliant random integer and
// indicate whether it is valid in CF.
RDSEED,
- // PCMP*STRI
PCMPISTRI,
PCMPESTRI,
- // XTEST - Test if in transactional execution.
+ // Test if in transactional execution.
XTEST,
- // LCMPXCHG_DAG, LCMPXCHG8_DAG, LCMPXCHG16_DAG - Compare and swap.
+ // ERI instructions
+ RSQRT28, RCP28, EXP2,
+
+ // Compare and swap.
LCMPXCHG_DAG = ISD::FIRST_TARGET_MEMORY_OPCODE,
LCMPXCHG8_DAG,
LCMPXCHG16_DAG,
- // VZEXT_LOAD - Load, scalar_to_vector, and zero extend.
+ // Load, scalar_to_vector, and zero extend.
VZEXT_LOAD,
- // FNSTCW16m - Store FP control world into i16 memory.
+ // Store FP control world into i16 memory.
FNSTCW16m,
- /// FP_TO_INT*_IN_MEM - This instruction implements FP_TO_SINT with the
+ /// This instruction implements FP_TO_SINT with the
/// integer destination in memory and a FP reg source. This corresponds
/// to the X86::FIST*m instructions and the rounding mode change stuff. It
/// has two inputs (token chain and address) and two outputs (int value
FP_TO_INT32_IN_MEM,
FP_TO_INT64_IN_MEM,
- /// FILD, FILD_FLAG - This instruction implements SINT_TO_FP with the
+ /// This instruction implements SINT_TO_FP with the
/// integer source in memory and FP reg result. This corresponds to the
/// X86::FILD*m instructions. It has three inputs (token chain, address,
/// and source type) and two outputs (FP value and token chain). FILD_FLAG
FILD,
FILD_FLAG,
- /// FLD - This instruction implements an extending load to FP stack slots.
+ /// This instruction implements an extending load to FP stack slots.
/// This corresponds to the X86::FLD32m / X86::FLD64m. It takes a chain
/// operand, ptr to load from, and a ValueType node indicating the type
/// to load to.
FLD,
- /// FST - This instruction implements a truncating store to FP stack
+ /// This instruction implements a truncating store to FP stack
/// slots. This corresponds to the X86::FST32m / X86::FST64m. It takes a
/// chain operand, value to store, address, and a ValueType to store it
/// as.
FST,
- /// VAARG_64 - This instruction grabs the address of the next argument
+ /// This instruction grabs the address of the next argument
/// from a va_list. (reads and modifies the va_list in memory)
VAARG_64
/// Define some predicates that are used for node matching.
namespace X86 {
- /// isVEXTRACT128Index - Return true if the specified
+ /// Return true if the specified
/// EXTRACT_SUBVECTOR operand specifies a vector extract that is
/// suitable for input to VEXTRACTF128, VEXTRACTI128 instructions.
bool isVEXTRACT128Index(SDNode *N);
- /// isVINSERT128Index - Return true if the specified
+ /// Return true if the specified
/// INSERT_SUBVECTOR operand specifies a subvector insert that is
/// suitable for input to VINSERTF128, VINSERTI128 instructions.
bool isVINSERT128Index(SDNode *N);
- /// isVEXTRACT256Index - Return true if the specified
+ /// Return true if the specified
/// EXTRACT_SUBVECTOR operand specifies a vector extract that is
/// suitable for input to VEXTRACTF64X4, VEXTRACTI64X4 instructions.
bool isVEXTRACT256Index(SDNode *N);
- /// isVINSERT256Index - Return true if the specified
+ /// Return true if the specified
/// INSERT_SUBVECTOR operand specifies a subvector insert that is
/// suitable for input to VINSERTF64X4, VINSERTI64X4 instructions.
bool isVINSERT256Index(SDNode *N);
- /// getExtractVEXTRACT128Immediate - Return the appropriate
+ /// Return the appropriate
/// immediate to extract the specified EXTRACT_SUBVECTOR index
/// with VEXTRACTF128, VEXTRACTI128 instructions.
unsigned getExtractVEXTRACT128Immediate(SDNode *N);
- /// getInsertVINSERT128Immediate - Return the appropriate
+ /// Return the appropriate
/// immediate to insert at the specified INSERT_SUBVECTOR index
/// with VINSERTF128, VINSERT128 instructions.
unsigned getInsertVINSERT128Immediate(SDNode *N);
- /// getExtractVEXTRACT256Immediate - Return the appropriate
+ /// Return the appropriate
/// immediate to extract the specified EXTRACT_SUBVECTOR index
/// with VEXTRACTF64X4, VEXTRACTI64x4 instructions.
unsigned getExtractVEXTRACT256Immediate(SDNode *N);
- /// getInsertVINSERT256Immediate - Return the appropriate
+ /// Return the appropriate
/// immediate to insert at the specified INSERT_SUBVECTOR index
/// with VINSERTF64x4, VINSERTI64x4 instructions.
unsigned getInsertVINSERT256Immediate(SDNode *N);
- /// isZeroNode - Returns true if Elt is a constant zero or a floating point
- /// constant +0.0.
+ /// Returns true if Elt is a constant zero or floating point constant +0.0.
bool isZeroNode(SDValue Elt);
- /// isOffsetSuitableForCodeModel - Returns true of the given offset can be
+ /// Returns true of the given offset can be
/// fit into displacement field of the instruction.
bool isOffsetSuitableForCodeModel(int64_t Offset, CodeModel::Model M,
bool hasSymbolicDisplacement = true);
- /// isCalleePop - Determines whether the callee is required to pop its
+ /// Determines whether the callee is required to pop its
/// own arguments. Callee pop is necessary to support tail calls.
bool isCalleePop(CallingConv::ID CallingConv,
bool is64Bit, bool IsVarArg, bool TailCallOpt);
}
//===--------------------------------------------------------------------===//
- // X86TargetLowering - X86 Implementation of the TargetLowering interface
+ // X86 Implementation of the TargetLowering interface
class X86TargetLowering final : public TargetLowering {
public:
- explicit X86TargetLowering(X86TargetMachine &TM);
+ explicit X86TargetLowering(const X86TargetMachine &TM,
+ const X86Subtarget &STI);
unsigned getJumpTableEncoding() const override;
+ bool useSoftFloat() const override;
- MVT getScalarShiftAmountTy(EVT LHSTy) const override { return MVT::i8; }
+ MVT getScalarShiftAmountTy(const DataLayout &, EVT) const override {
+ return MVT::i8;
+ }
const MCExpr *
LowerCustomJumpTableEntry(const MachineJumpTableInfo *MJTI,
const MachineBasicBlock *MBB, unsigned uid,
MCContext &Ctx) const override;
- /// getPICJumpTableRelocaBase - Returns relocation base for the given PIC
- /// jumptable.
+ /// Returns relocation base for the given PIC jumptable.
SDValue getPICJumpTableRelocBase(SDValue Table,
SelectionDAG &DAG) const override;
const MCExpr *
getPICJumpTableRelocBaseExpr(const MachineFunction *MF,
unsigned JTI, MCContext &Ctx) const override;
- /// getByValTypeAlignment - Return the desired alignment for ByVal aggregate
+ /// Return the desired alignment for ByVal aggregate
/// function arguments in the caller parameter area. For X86, aggregates
/// that contains are placed at 16-byte boundaries while the rest are at
/// 4-byte boundaries.
- unsigned getByValTypeAlignment(Type *Ty) const override;
+ unsigned getByValTypeAlignment(Type *Ty,
+ const DataLayout &DL) const override;
- /// getOptimalMemOpType - Returns the target specific optimal type for load
+ /// Returns the target specific optimal type for load
/// and store operations as a result of memset, memcpy, and memmove
/// lowering. If DstAlign is zero that means it's safe to destination
/// alignment can satisfy any constraint. Similarly if SrcAlign is zero it
bool IsMemset, bool ZeroMemset, bool MemcpyStrSrc,
MachineFunction &MF) const override;
- /// isSafeMemOpType - Returns true if it's safe to use load / store of the
+ /// Returns true if it's safe to use load / store of the
/// specified type to expand memcpy / memset inline. This is mostly true
/// for all types except for some special cases. For example, on X86
/// targets without SSE2 f64 load / store are done with fldl / fstpl which
/// legal as the hook is used before type legalization.
bool isSafeMemOpType(MVT VT) const override;
- /// allowsMisalignedMemoryAccesses - Returns true if the target allows
- /// unaligned memory accesses. of the specified type. Returns whether it
- /// is "fast" by reference in the second argument.
+ /// Returns true if the target allows unaligned memory accesses of the
+ /// specified type. Returns whether it is "fast" in the last argument.
bool allowsMisalignedMemoryAccesses(EVT VT, unsigned AS, unsigned Align,
bool *Fast) const override;
- /// LowerOperation - Provide custom lowering hooks for some operations.
+ /// Provide custom lowering hooks for some operations.
///
SDValue LowerOperation(SDValue Op, SelectionDAG &DAG) const override;
- /// ReplaceNodeResults - Replace the results of node with an illegal result
+ /// Replace the results of node with an illegal result
/// type with new values built out of custom code.
///
void ReplaceNodeResults(SDNode *N, SmallVectorImpl<SDValue>&Results,
SDValue PerformDAGCombine(SDNode *N, DAGCombinerInfo &DCI) const override;
- /// isTypeDesirableForOp - Return true if the target has native support for
+ /// Return true if the target has native support for
/// the specified value type and it is 'desirable' to use the type for the
/// given node type. e.g. On x86 i16 is legal, but undesirable since i16
/// instruction encodings are longer and some i16 instructions are slow.
bool isTypeDesirableForOp(unsigned Opc, EVT VT) const override;
- /// isTypeDesirable - Return true if the target has native support for the
+ /// Return true if the target has native support for the
/// specified value type and it is 'desirable' to use the type. e.g. On x86
/// i16 is legal, but undesirable since i16 instruction encodings are longer
/// and some i16 instructions are slow.
MachineBasicBlock *MBB) const override;
- /// getTargetNodeName - This method returns the name of a target specific
- /// DAG node.
+ /// This method returns the name of a target specific DAG node.
const char *getTargetNodeName(unsigned Opcode) const override;
- /// getSetCCResultType - Return the value type to use for ISD::SETCC.
- EVT getSetCCResultType(LLVMContext &Context, EVT VT) const override;
+ bool isCheapToSpeculateCttz() const override;
+
+ bool isCheapToSpeculateCtlz() const override;
- /// computeKnownBitsForTargetNode - Determine which of the bits specified
- /// in Mask are known to be either zero or one and return them in the
- /// KnownZero/KnownOne bitsets.
+ /// Return the value type to use for ISD::SETCC.
+ EVT getSetCCResultType(const DataLayout &DL, LLVMContext &Context,
+ EVT VT) const override;
+
+ /// Determine which of the bits specified in Mask are known to be either
+ /// zero or one and return them in the KnownZero/KnownOne bitsets.
void computeKnownBitsForTargetNode(const SDValue Op,
APInt &KnownZero,
APInt &KnownOne,
const SelectionDAG &DAG,
unsigned Depth = 0) const override;
- // ComputeNumSignBitsForTargetNode - Determine the number of bits in the
- // operation that are sign bits.
+ /// Determine the number of bits in the operation that are sign bits.
unsigned ComputeNumSignBitsForTargetNode(SDValue Op,
const SelectionDAG &DAG,
unsigned Depth) const override;
bool ExpandInlineAsm(CallInst *CI) const override;
- ConstraintType
- getConstraintType(const std::string &Constraint) const override;
+ ConstraintType getConstraintType(StringRef Constraint) const override;
/// Examine constraint string and operand type and determine a weight value.
/// The operand object must already have been set up with the operand type.
const char *LowerXConstraint(EVT ConstraintVT) const override;
- /// LowerAsmOperandForConstraint - Lower the specified operand into the Ops
- /// vector. If it is invalid, don't add anything to Ops. If hasMemory is
- /// true it means one of the asm constraint of the inline asm instruction
- /// being processed is 'm'.
+ /// Lower the specified operand into the Ops vector. If it is invalid, don't
+ /// add anything to Ops. If hasMemory is true it means one of the asm
+ /// constraint of the inline asm instruction being processed is 'm'.
void LowerAsmOperandForConstraint(SDValue Op,
std::string &Constraint,
std::vector<SDValue> &Ops,
SelectionDAG &DAG) const override;
- /// getRegForInlineAsmConstraint - Given a physical register constraint
+ unsigned
+ getInlineAsmMemConstraint(StringRef ConstraintCode) const override {
+ if (ConstraintCode == "i")
+ return InlineAsm::Constraint_i;
+ else if (ConstraintCode == "o")
+ return InlineAsm::Constraint_o;
+ else if (ConstraintCode == "v")
+ return InlineAsm::Constraint_v;
+ else if (ConstraintCode == "X")
+ return InlineAsm::Constraint_X;
+ return TargetLowering::getInlineAsmMemConstraint(ConstraintCode);
+ }
+
+ /// Given a physical register constraint
/// (e.g. {edx}), return the register number and the register class for the
/// register. This should only be used for C_Register constraints. On
/// error, this returns a register number of 0.
- std::pair<unsigned, const TargetRegisterClass*>
- getRegForInlineAsmConstraint(const std::string &Constraint,
- MVT VT) const override;
+ std::pair<unsigned, const TargetRegisterClass *>
+ getRegForInlineAsmConstraint(const TargetRegisterInfo *TRI,
+ StringRef Constraint, MVT VT) const override;
- /// isLegalAddressingMode - Return true if the addressing mode represented
+ /// Return true if the addressing mode represented
/// by AM is legal for this target, for a load/store of the specified type.
- bool isLegalAddressingMode(const AddrMode &AM, Type *Ty) const override;
+ bool isLegalAddressingMode(const DataLayout &DL, const AddrMode &AM,
+ Type *Ty, unsigned AS) const override;
- /// isLegalICmpImmediate - Return true if the specified immediate is legal
+ /// Return true if the specified immediate is legal
/// icmp immediate, that is the target has icmp instructions which can
/// compare a register against the immediate without having to materialize
/// the immediate into a register.
bool isLegalICmpImmediate(int64_t Imm) const override;
- /// isLegalAddImmediate - Return true if the specified immediate is legal
+ /// Return true if the specified immediate is legal
/// add immediate, that is the target has add instructions which can
/// add a register and the immediate without having to materialize
/// the immediate into a register.
/// of the specified type.
/// If the AM is supported, the return value must be >= 0.
/// If the AM is not supported, it returns a negative value.
- int getScalingFactorCost(const AddrMode &AM, Type *Ty) const override;
+ int getScalingFactorCost(const DataLayout &DL, const AddrMode &AM, Type *Ty,
+ unsigned AS) const override;
bool isVectorShiftByScalarCheap(Type *Ty) const override;
- /// isTruncateFree - Return true if it's free to truncate a value of
+ /// Return true if it's free to truncate a value of
/// type Ty1 to type Ty2. e.g. On x86 it's free to truncate a i32 value in
/// register EAX to i16 by referencing its sub-register AX.
bool isTruncateFree(Type *Ty1, Type *Ty2) const override;
bool allowTruncateForTailCall(Type *Ty1, Type *Ty2) const override;
- /// isZExtFree - Return true if any actual instruction that defines a
+ /// Return true if any actual instruction that defines a
/// value of type Ty1 implicit zero-extends the value to Ty2 in the result
/// register. This does not necessarily include registers defined in
/// unknown ways, such as incoming arguments, or copies from unknown
bool isZExtFree(EVT VT1, EVT VT2) const override;
bool isZExtFree(SDValue Val, EVT VT2) const override;
- /// isFMAFasterThanFMulAndFAdd - Return true if an FMA operation is faster
- /// than a pair of fmul and fadd instructions. fmuladd intrinsics will be
- /// expanded to FMAs when this method returns true, otherwise fmuladd is
- /// expanded to fmul + fadd.
+ /// Return true if folding a vector load into ExtVal (a sign, zero, or any
+ /// extend node) is profitable.
+ bool isVectorLoadExtDesirable(SDValue) const override;
+
+ /// Return true if an FMA operation is faster than a pair of fmul and fadd
+ /// instructions. fmuladd intrinsics will be expanded to FMAs when this
+ /// method returns true, otherwise fmuladd is expanded to fmul + fadd.
bool isFMAFasterThanFMulAndFAdd(EVT VT) const override;
- /// isNarrowingProfitable - Return true if it's profitable to narrow
+ /// Return true if it's profitable to narrow
/// operations of type VT1 to VT2. e.g. on x86, it's profitable to narrow
/// from i32 to i8 but not from i32 to i16.
bool isNarrowingProfitable(EVT VT1, EVT VT2) const override;
- /// isFPImmLegal - Returns true if the target can instruction select the
+ /// Returns true if the target can instruction select the
/// specified FP immediate natively. If false, the legalizer will
/// materialize the FP immediate as a load from a constant pool.
bool isFPImmLegal(const APFloat &Imm, EVT VT) const override;
- /// isShuffleMaskLegal - Targets can use this to indicate that they only
- /// support *some* VECTOR_SHUFFLE operations, those with specific masks.
- /// By default, if a target supports the VECTOR_SHUFFLE node, all mask
- /// values are assumed to be legal.
+ /// Targets can use this to indicate that they only support *some*
+ /// VECTOR_SHUFFLE operations, those with specific masks. By default, if a
+ /// target supports the VECTOR_SHUFFLE node, all mask values are assumed to
+ /// be legal.
bool isShuffleMaskLegal(const SmallVectorImpl<int> &Mask,
EVT VT) const override;
- /// isVectorClearMaskLegal - Similar to isShuffleMaskLegal. This is
- /// used by Targets can use this to indicate if there is a suitable
- /// VECTOR_SHUFFLE that can be used to replace a VAND with a constant
- /// pool entry.
+ /// Similar to isShuffleMaskLegal. This is used by Targets can use this to
+ /// indicate if there is a suitable VECTOR_SHUFFLE that can be used to
+ /// replace a VAND with a constant pool entry.
bool isVectorClearMaskLegal(const SmallVectorImpl<int> &Mask,
EVT VT) const override;
- /// ShouldShrinkFPConstant - If true, then instruction selection should
+ /// If true, then instruction selection should
/// seek to shrink the FP constant of the specified type to a smaller type
/// in order to save space and / or reduce runtime.
bool ShouldShrinkFPConstant(EVT VT) const override {
return !X86ScalarSSEf64 || VT == MVT::f80;
}
- const X86Subtarget* getSubtarget() const {
- return Subtarget;
- }
+ /// Return true if we believe it is correct and profitable to reduce the
+ /// load node to a smaller type.
+ bool shouldReduceLoadWidth(SDNode *Load, ISD::LoadExtType ExtTy,
+ EVT NewVT) const override;
- /// isScalarFPTypeInSSEReg - Return true if the specified scalar FP type is
- /// computed in an SSE register, not on the X87 floating point stack.
+ /// Return true if the specified scalar FP type is computed in an SSE
+ /// register, not on the X87 floating point stack.
bool isScalarFPTypeInSSEReg(EVT VT) const {
return (VT == MVT::f64 && X86ScalarSSEf64) || // f64 is when SSE2
(VT == MVT::f32 && X86ScalarSSEf32); // f32 is when SSE1
}
- /// isTargetFTOL - Return true if the target uses the MSVC _ftol2 routine
- /// for fptoui.
- bool isTargetFTOL() const;
-
- /// isIntegerTypeFTOL - Return true if the MSVC _ftol2 routine should be
- /// used for fptoui to the given type.
- bool isIntegerTypeFTOL(EVT VT) const {
- return isTargetFTOL() && VT == MVT::i64;
- }
-
/// \brief Returns true if it is beneficial to convert a load of a constant
/// to just the constant itself.
bool shouldConvertConstantLoadToIntImm(const APInt &Imm,
Type *Ty) const override;
+ /// Return true if EXTRACT_SUBVECTOR is cheap for this result type
+ /// with this index.
+ bool isExtractSubvectorCheap(EVT ResVT, unsigned Index) const override;
+
/// Intel processors have a unified instruction and data cache
const char * getClearCacheBuiltinName() const override {
return nullptr; // nothing to do, move along.
}
- unsigned getRegisterByName(const char* RegName, EVT VT) const override;
+ unsigned getRegisterByName(const char* RegName, EVT VT,
+ SelectionDAG &DAG) const override;
- /// createFastISel - This method returns a target specific FastISel object,
+ /// This method returns a target specific FastISel object,
/// or null if the target does not support "fast" ISel.
FastISel *createFastISel(FunctionLoweringInfo &funcInfo,
const TargetLibraryInfo *libInfo) const override;
- /// getStackCookieLocation - Return true if the target stores stack
- /// protector cookies at a fixed offset in some non-standard address
- /// space, and populates the address space and offset as
- /// appropriate.
+ /// Return true if the target stores stack protector cookies at a fixed
+ /// offset in some non-standard address space, and populates the address
+ /// space and offset as appropriate.
bool getStackCookieLocation(unsigned &AddressSpace,
unsigned &Offset) const override;
+ /// Return true if the target stores SafeStack pointer at a fixed offset in
+ /// some non-standard address space, and populates the address space and
+ /// offset as appropriate.
+ bool getSafeStackPointerLocation(unsigned &AddressSpace,
+ unsigned &Offset) const override;
+
SDValue BuildFILD(SDValue Op, EVT SrcVT, SDValue Chain, SDValue StackSlot,
SelectionDAG &DAG) const;
bool isNoopAddrSpaceCast(unsigned SrcAS, unsigned DestAS) const override;
- /// \brief Reset the operation actions based on target options.
- void resetOperationActions() override;
-
bool useLoadStackGuardNode() const override;
/// \brief Customize the preferred legalization strategy for certain types.
LegalizeTypeAction getPreferredVectorAction(EVT VT) const override;
+ bool isIntDivCheap(EVT VT, AttributeSet Attr) const override;
+
protected:
- std::pair<const TargetRegisterClass*, uint8_t>
- findRepresentativeClass(MVT VT) const override;
+ std::pair<const TargetRegisterClass *, uint8_t>
+ findRepresentativeClass(const TargetRegisterInfo *TRI,
+ MVT VT) const override;
private:
- /// Subtarget - Keep a pointer to the X86Subtarget around so that we can
+ /// Keep a pointer to the X86Subtarget around so that we can
/// make the right decision when generating code for different targets.
const X86Subtarget *Subtarget;
const DataLayout *TD;
- /// Used to store the TargetOptions so that we don't waste time resetting
- /// the operation actions unless we have to.
- TargetOptions TO;
-
- /// X86ScalarSSEf32, X86ScalarSSEf64 - Select between SSE or x87
- /// floating point ops.
+ /// Select between SSE or x87 floating point ops.
/// When SSE is available, use it for f32 operations.
/// When SSE2 is available, use it for f64 operations.
bool X86ScalarSSEf32;
bool X86ScalarSSEf64;
- /// LegalFPImmediates - A list of legal fp immediates.
+ /// A list of legal FP immediates.
std::vector<APFloat> LegalFPImmediates;
- /// addLegalFPImmediate - Indicate that this x86 target can instruction
+ /// Indicate that this x86 target can instruction
/// select the specified FP immediate natively.
void addLegalFPImmediate(const APFloat& Imm) {
LegalFPImmediates.push_back(Imm);
// Call lowering helpers.
- /// IsEligibleForTailCallOptimization - Check whether the call is eligible
- /// for tail call optimization. Targets which want to do tail call
- /// optimization should implement this function.
+ /// Check whether the call is eligible for tail call optimization. Targets
+ /// that want to do tail call optimization should implement this function.
bool IsEligibleForTailCallOptimization(SDValue Callee,
CallingConv::ID CalleeCC,
bool isVarArg,
SDValue LowerFRAMEADDR(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerFRAME_TO_ARGS_OFFSET(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerEH_RETURN(SDValue Op, SelectionDAG &DAG) const;
+ SDValue LowerCLEANUPRET(SDValue Op, SelectionDAG &DAG) const;
SDValue lowerEH_SJLJ_SETJMP(SDValue Op, SelectionDAG &DAG) const;
SDValue lowerEH_SJLJ_LONGJMP(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerINIT_TRAMPOLINE(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerFLT_ROUNDS_(SDValue Op, SelectionDAG &DAG) const;
- SDValue LowerSIGN_EXTEND_INREG(SDValue Op, SelectionDAG &DAG) const;
SDValue LowerWin64_i128OP(SDValue Op, SelectionDAG &DAG) const;
+ SDValue LowerGC_TRANSITION_START(SDValue Op, SelectionDAG &DAG) const;
+ SDValue LowerGC_TRANSITION_END(SDValue Op, SelectionDAG &DAG) const;
SDValue
LowerFormalArguments(SDValue Chain,
const MCPhysReg *getScratchRegisters(CallingConv::ID CC) const override;
- bool shouldExpandAtomicLoadInIR(LoadInst *SI) const override;
+ TargetLoweringBase::AtomicExpansionKind
+ shouldExpandAtomicLoadInIR(LoadInst *SI) const override;
bool shouldExpandAtomicStoreInIR(StoreInst *SI) const override;
- bool shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const override;
+ TargetLoweringBase::AtomicExpansionKind
+ shouldExpandAtomicRMWInIR(AtomicRMWInst *AI) const override;
+
+ LoadInst *
+ lowerIdempotentRMWIntoFencedLoad(AtomicRMWInst *AI) const override;
- bool needsCmpXchgNb(const Type *MemType) const;
+ bool needsCmpXchgNb(Type *MemType) const;
/// Utility function to emit atomic-load-arith operations (and, or, xor,
/// nand, max, min, umax, umin). It takes the corresponding instruction to
MachineBasicBlock *EmitLoweredSelect(MachineInstr *I,
MachineBasicBlock *BB) const;
+ MachineBasicBlock *EmitLoweredAtomicFP(MachineInstr *I,
+ MachineBasicBlock *BB) const;
+
MachineBasicBlock *EmitLoweredWinAlloca(MachineInstr *MI,
MachineBasicBlock *BB) const;
/// Convert a comparison if required by the subtarget.
SDValue ConvertCmpIfNecessary(SDValue Cmp, SelectionDAG &DAG) const;
+
+ /// Use rsqrt* to speed up sqrt calculations.
+ SDValue getRsqrtEstimate(SDValue Operand, DAGCombinerInfo &DCI,
+ unsigned &RefinementSteps,
+ bool &UseOneConstNR) const override;
+
+ /// Use rcp* to speed up fdiv calculations.
+ SDValue getRecipEstimate(SDValue Operand, DAGCombinerInfo &DCI,
+ unsigned &RefinementSteps) const override;
+
+ /// Reassociate floating point divisions into multiply by reciprocal.
+ unsigned combineRepeatedFPDivisors() const override;
};
namespace X86 {